Introduction - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 1

1 Introduction

Contributors: G. Blöschl,* M. Sivapalan, T. Wagener,
A. Viglione and H. H. G. Savenije

1.1 Why we need runoff predictions
During the February 2007 Zambezi River flood, Paulo
Zucula, the director of Mozambique

failure. What makes the operation of the dam even more
complicated is that large parts of the upstream catchment
of the Zambezi are literally ungauged. Runoff in the main
stream of the Zambezi River may be governed by the
upstream Kariba Dam, from where warnings are issued
whenever they open the flood gates, but the operators have
no knowledge of the inflow from the intermediate catch-
ment, of which the Luangwa with its 50 000 km 2 is the
largest. The Luangwa is completely ungauged. As a result,
operators sometimes have to open the floodgates and
discharge more water than would be necessary, with the
benefit of hindsight. Better runoff predictions in the
Luangwa could reduce flood releases, increase hydropower
production, improve flood warning, and reduce down-
stream damage and suffering.
In Chapter 11 , a case study by Hessel Winsemius shows
that, even in an ungauged basin such as the Luangwa,
much can be done in terms of improved flood predictions.
Figure 1.2 is a screen dump of the online model that
Winsemius developed for the Luangwa River basin, com-
pletely based on remotely sensed data (mostly precipitation
and meteorological

s National Institute
for Disaster Management, was trying to contain the disas-
ter along the river:

'

The evacuated people have been in
camps for over a week without proper feeding

'

… they are

isolated and we can t go there by road, so we have to airlift

some of them and drop food,

'

'

he said. Some 90 000 people
were made homeless by the flood. According to an Oxfam
worker, about 1000 people a day were arriving at the
camps even without any shelter being provided. The gov-
ernment had learned the lessons of the previous 2001
flood, however, during which about 700 people died. This
time it promptly launched missions by boat and helicopter
to evacuate people from affected areas. But they were
rapidly running short of food for the people in the 33
temporary camps, which also lacked tents, medicine and
clean water.
In January 2008 a major flood again struck the Zambezi.
This time some 50 000 people in Mozambique were dis-
placed by the flood. Property and infrastructure is again

being wrecked but we are more worried about the people,

'

information). The model has been

'
said Paulo Zucula. The flood in the Zambezi valley was in
fact worse than the floods of February 2007, and the
authorities were forced to evacuate areas where the victims
of earlier floods had been resettled.
What has this disaster got to do with runoff Predictions
in Ungauged Basins (PUB)? A lot! The hydrology of the
Zambezi valley in Mozambique is strongly affected by the
presence of the Cahora Bassa Dam (see Figure 1.1 ). This
dam, designed to release about 1900 m 3 /s through its
turbines for hydropower generation, has a limited flood
release capacity. In order to deal with major flooding, it
has to lower its reservoir level substantially before the
onset of the flood season each year. It is a perpetual
trade-off between the economic value of hydropower gen-
eration, the risk of flood damage and the risk of dam

Figure 1.1. The Cahora Bassa Dam, Mozambique, spilling through
one of the eight flood gates.